causation

1 Setup

1.1 load packages

1.2 read data and set variables

2 check N

## # A tibble: 1 × 1
##       n
##   <int>
## 1    50

3 Freesort: Plotting empirical euclidean distances

3.1 reshape data

3.2 plot subject-specific raw responses

3.3 plot average raw responses

3.4 plot subject-specific euclidean rdms

3.5 plot average euclidean rdm

4 Causal

4.1 reshape data

4.2 plot subject-specific causal RDMs

4.3 plot average causal RDM

5 Correlations: Empirical euclidean RDMs

5.1 read and compute correlations

5.2 plot correlations with noise ceiling

6 Correlations: Causal RDMS

6.1 generate causal theories

d_theo_causal_raw <- read_csv(here("code/analysis/pilot/theoretical-rdm-data/d_theories_categories.csv"))

d_theo_causal <- d_theo_causal_raw %>% 
  select(itemA, itemB, MindBodyAction_itemA, MindBodyAction_itemB) %>% 
  mutate(within_domain_int = case_when(
                          #intuitive bio
                          MindBodyAction_itemA == "bio" & MindBodyAction_itemB == "act" ~ 0, 
                          MindBodyAction_itemA == "act" & MindBodyAction_itemB == "bio" ~ 0,
                          MindBodyAction_itemA == "bio" & MindBodyAction_itemB == "bio" ~ 0,
                          #intuitive psych
                          MindBodyAction_itemA == "mind" & MindBodyAction_itemB == "act" ~ 0, 
                          MindBodyAction_itemA == "act" & MindBodyAction_itemB == "mind" ~ 0,
                          MindBodyAction_itemA == "mind" & MindBodyAction_itemB == "mind" ~ 0,
                          TRUE ~ 1),
         cross_domain_int = case_when(
                          MindBodyAction_itemA == "bio" & MindBodyAction_itemB == "act" ~ 0, 
                          MindBodyAction_itemA == "act" & MindBodyAction_itemB == "bio" ~ 0,
                          MindBodyAction_itemA == "bio" & MindBodyAction_itemB == "bio" ~ 0,
                          
                          MindBodyAction_itemA == "mind" & MindBodyAction_itemB == "act" ~ 0, 
                          MindBodyAction_itemA == "act" & MindBodyAction_itemB == "mind" ~ 0,
                          MindBodyAction_itemA == "mind" & MindBodyAction_itemB == "mind" ~ 0,
                          #cross-domain causation
                          MindBodyAction_itemA == "mind" & MindBodyAction_itemB == "bio" ~ 0, 
                          MindBodyAction_itemA == "bio" & MindBodyAction_itemB == "mind" ~ 0,
                          TRUE ~ 1)) %>% 
  select(-c("MindBodyAction_itemA", "MindBodyAction_itemB"))

6.2 compute correlations

6.3 plot correlations with empirical causal rdms

key

1. _upper: the upper tri of the causal rdm (since causal isn’t symmetric, doing correlations for each causal tri separately. always using upper tri for empirical euclidean rdms since those are symmetric)
2. _lower: lower tri
   
3. _both: remove NAs, flatten into a vector, then compute correlations between the entire theoretical rdm and the causal rdms (NB: both is not avg of upper and lower)
   
4. within_domain_: a classical causal rdm involving intuitive theories within domains. mind causes act (and vice versa), bio causes act (and vice versa), mind causes mind, and bio causes bio.
   
5. cross_domain_: classical causal rdm per point 4 PLUS causal relation between bio and mind (bio cause mind and vice versa)

still to do: causal rdms where everything can cause everything except for certain conditions..

• There are limits to what can cause perception (seeing or hearing something) - at least directly. For instance, mental and bodily events can’t literally cause you to see something. But moving one’s body around can.
  
• Certain physiological events limit what the body can do - e.g. feeling sick and hungry tends to not lead to high-intensity actions like jumping, kicking, walking. Energy flows downhill.
  
• Certain actions constrain what next actions are possible, e.g. resting actions like sitting down tend not to cause high energy actions.
  
• ??? Perceptual events (seeing or hearing) cannot directly cause physiological states like being tired, being in pain - but cognitive events can.